• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2019, Vol. 55 ›› Issue (2): 91-97.doi: 10.3901/JME.2019.02.091

• 材料科学与工程 • 上一篇    下一篇

基于渐变凹模圆角半径的高强钢扭曲回弹补偿

谢延敏, 张飞, 王子豪, 黄仁勇, 杨俊峰, 胡云川   

  1. 西南交通大学机械工程学院 成都 610031
  • 收稿日期:2018-01-31 修回日期:2018-07-18 出版日期:2019-01-20 发布日期:2019-01-20
  • 通讯作者: 谢延敏(通信作者),男,1975年出生,博士,副教授,硕士研究生导师。主要研究方向为先进塑性加工技术仿真和稳健设计等。E-mail:xie_yanmin@swjtu.edu.cn
  • 基金资助:
    国家自然科学基金(51005193)和国家大学生创新性实验计划(201710613033)资助项目

Compensation of Twist Springback in High-strength Steel Based on Gradient Die Radius

XIE Yanmin, ZHANG Fei, WANG Zihao, HUANG Renyong, YANG Junfeng, HU Yunchuan   

  1. School of Mechanical Engineering, Southwest Jiaotong University, Chengdu 610031
  • Received:2018-01-31 Revised:2018-07-18 Online:2019-01-20 Published:2019-01-20

摘要: 为减少高强钢冲压成形扭曲回弹,提出一种基于渐变凹模圆角半径的模具补偿方法。以高强钢TRIP780双C件为研究对象,采用板料冲压成形仿真软件DYNAFORM对该双C件的冲压、扭曲回弹过程进行数值模拟。提出一种评价双C件扭曲回弹程度的指标,并进行双C件扭曲回弹试验,通过三坐标测量仪测量扭曲回弹角,对有限元模型进行了验证。以冲压成形后的扭曲回弹为优化目标,结合相关的工艺参数,利用BP神经网络,基于正交试验,建立凹模圆角半径渐变量、工艺参数与扭曲回弹角之间的网络模型。最后采用遗传算法对该模型迭代寻优获得最优凹模圆角半径渐变量和工艺参数。通过对比优化前后的扭曲回弹角,证明了优化流程有效地减少了双C件扭曲回弹。该方法为扭曲回弹的控制提供了一种新的思路。

关键词: BP神经网络, 高强钢, 渐变凹模圆角半径, 扭曲回弹, 遗传算法

Abstract: In order to reduce the twist springback appearing after the stamping of high-strength steel, a compensation method with gradient die radius is proposed. The double C rail of TRIP780 high-strength steel is taken as the research object. The sheet metal stamping simulation software DYNAFORM is used to numerically simulate the stamping and twist springback processes of the double C rail. An index to evaluate the twist springback of the double C rail is proposed. The experiment of twist springback for the double C rail is carried out. The twist springback angle is measured by means of a three-coordinate measuring instrument, and the finite element model is validated. The twist springback appearing after stamping is taken as the optimization target, and the related process parameters are taken into account. BP neural network is used to establish the network model between the gradient variation of die radius, process parameters and the twist springback angle based on orthogonal test. Finally, the model is iteratively optimized using genetic algorithm to obtain the optimum gradient variation of die radius and process parameters. The twist springback angle is compared before and after the optimization, which proves the optimization flow efficient to reduce the twist springback of the double C rail. This method provides a new way for the control of the twist springback.

Key words: BP neural network, genetic algorithm, gradient die radius, high-strength steel, twist springback

中图分类号: